Ttie objective of ttiis proposed study is to investigate ttie degree to wtiichi nnetfornnin, a lifestyle intervention, or both, can reduce breast cancer mortality among overweiqtit/obese, postmenopausal breast cancer survivors. We will use a "Biomarker Bridge" design that links clinical outcomes from a breast cancer survivor cohort with intermediate outcomes from a randomized controlled trial by means of a Biomarker Risk Score. Biomarker Risk Score Development: We will assay panels of interrelated biomarkers in 375 archived blood samples obtained from oven/veight/obese, postmenopausal women with a history of breast cancer (125 cases [breast cancer death]:250 matched controls). These Biomarkers represent proposed mechanisms by which obesity is associated with postmenopausal breast cancer: (1) alterations in the insulin-IGF axis, (2) concentrations of endogenous sex hormones, and (3) chronic inflammation. We will identify a set of markers that best predicts breast cancer mortality using conditional logistic regression models adjusted for prognostic factors. This model (i.e., Biomarker Risk Score) measures the log-odds of disease risk due to joint biomarker concentrations. Therefore this Risk Score can measure changes in log-odds due to changes in these markers in an individual and will be used to assess the clinical impact of the metformin and lifestyle intervention randomized trial. Metformin/Lifestyle Intervention Trial: We will conduct a 6-month, randomized controlled trial in 340 overweight/obese, postmenopausal breast cancer survivors. Participants will be randomized in equal numbers to (1) placebo, (2) metformin, (3) lifestyle intervention and placebo, or (4) lifestyle intervention and metformin. The lifestyle intervention will focus on reducing energy intake and increasing energy expenditure to achieve a 7% weight reduction. Biomarkers that compose the Risk Score will be assayed in fasting blood samples collected at baseline and 6 months. The degree to which each intervention changes (e.g., reduces) the Biomarker Risk Score will be used to predict changes (e.g., reductions) in breast cancer mortality. We hypothesize that metformin and lifestyle interventions will reduce breast cancer mortality and that the combination of those interventions will have an additive effect on lowering nsk. In summary, this Biomarker Bridge Design will (i) develop a Biomarker Risk Score that predicts breast cancer mortality, (ii) examine how a metformin/lifestyle intervention changes this Risk Score, and (iii) thereby assess the degree to which metformin and lifestyle interventions influence the biological processes linking obesity with breast cancer mortality.